A new suicide risk prediction algorithm that targets soldiers at high risk of suicide after hospitalization for psychiatric treatment might help target preventive interventions, according to the Army Study to Assess Risk and Resilience in Servicemembers, or Army STARRS study.

In total, 52.9% of the posthospitalization suicides occurred after the 5% of hospitalizations with the highest predicted suicide risk (3,824.1 suicides per 100,000 person-years), reported Ronald C. Kessler, Ph.D., of Harvard Medical School, Boston, and his colleagues in a study published online (JAMA Psychiatry 2014 Nov. 12 [ doi: 10.1001/jamapsychiatry.2014.1754 ]). The U.S. Army suicide rate has increased since 2004 and exceeds the civilian rate despite numerous prevention and intervention programs, the study authors noted.

One potentially important group for targeted intervention is soldiers recently discharged from inpatient psychiatric treatment, who according to military documents have an eightfold elevated suicide risk in the 3 months after psychiatric hospitalization and a fivefold elevated risk for the 12 months after discharge.

A promising approach to assessing suicide risk in this group would be to use administrative data available during hospitalization to generate an actuarial posthospitalization suicide risk algorithm, the investigators said. “Previous research has revealed that actuarial suicide prediction is much more accurate than prediction based on clinical judgment,” they wrote. Using a range of administrative data, the researchers documented 53,769 hospitalizations of active-duty soldiers from January 2004 to December 2009 with psychiatric admission diagnoses.

A total of 68 soldiers died by suicide within 12 months of being discharged from hospital (12% of all U.S. Army suicides), equivalent to 263.9 suicides per 100,000 person-years, compared with 18.5 suicides per 100,000 person-years in the total U.S. Army.

Soldiers in the highest predicted suicide risk group had seven unintentional injury deaths, 830 suicide attempts, and 3,765 subsequent hospitalizations within 12 months of hospital discharge. The strongest predictors for suicide were male gender (odds ratio, 7.9; 95% confidence interval, 1.9-32.6); late age at enlistment (OR, 1.9; 95% CI, 1.0-3.5); and criminal offenses such as verbal violence (OR, 2.2; 95% CI, 1.2-4.0) and weapons possession (OR, 5.6; 95% CI, 1.7-18.3).

Other predictors included prior suicidality (OR, 2.9; 95% CI, 1.7-4.9) and prior psychiatric treatment, such as the number of prescriptions filled in the past 12 months (OR, 1.3; 95% CI, 1.1-1.7).

“Although interventions in this high-risk stratum would not solve the entire U.S. Army suicide problem, given that all posthospitalization suicides account for 12% of all U.S. Army suicides, the algorithm would presumably help target preventive interventions,” the investigators wrote. However, before the algorithm was used in clinical practice, several issues needed to be addressed, such as whether the algorithm was sufficiently stable to predict future suicides given that it was based on 68 suicides.

Also, one concern is that intensive posthospitalization interventions could lead to undue nonmedical scrutiny that might have detrimental effects on soldiers’ careers. “This concern is all the more important given that most soldiers identified as being high risk do not commit suicide,” Dr. Kessler and his colleagues wrote.

No authors made conflict of interest disclosures. The Army STARRS was sponsored by the U.S. Department of the Army and funded by the National Institute of Mental Health, National Institutes of Health, U.S. Department of Health and Human Services.